Quoting: slaapliedjeSteamOS is better than Ubuntu, for a game console. There are thongs they tune for the kernel.

And I have to ask.. do you have gorilla hands? I wouls say (with the exception of the Atari Jaguar controller) the Steam Controller has the largest hand grips out of any I have seen. The Nintendo Wireless Pro is a close second. This coming from someone who can't use the PS3 controller for very long because of how small they are.

Any distribution can have a tuned kernel, provided you know what hardware you are tuning for. Mine is rack-mounted; so not exactly what SteamOS is intended for.

As to gorilla hands, I am 2m in height, with all other bits of me in proportion. Meaning each of my hands is longer than the the steam controller is wide; I also require quite large thongs, although I have always known tham as flip-flops :D

View this comment - View article - View full comments

Quoting: papajo

Quoting: etonbears

Quoting: tonR

Quoting: NeverthelessI don't think it ever was a device species at all. I don't know what Valve thought of it when they released it, but for me it was a Linux PC with a controller for the living room. That concept is not dead, nor are the Steam Machines sold. Nothing has failed here, because nothing gets abandoned. Owners of a Steam Machine still have got a Linux PC for the living room and 4000+ games to play.

That's my question. What was/is in Valve's people mind when they launched Steam Machine? To promoted Linux? For fun?

I never said failed. Yes, Valve is still making money, Steam Machine or not, but every businesses in this world wanted to maximizing profit to even every single cent. So I'm very curious about the reason of Steam Machine existance.

I don't think there was a clear idea, on Valve's part, what a Steam Machine was for. It was a strand in the "open" gaming platform they wanted to exist, just as SteamOS and the Linux Steam runtime are strands of that platform.

My recollection is that they stated that they wished to foster hardware innovation, so there was no fixed/clear definition of a Steam Machine, but that they expected novel and upgradeable hardware offerings ( not just living room PCs ).

Unfortunately, that is difficult to achieve. PCs are already, de facto, flexible and upgradeable, so what can be added to that? Novel hardware that might have superior performance or price characteristics would take the form of custom designed components; but that requires a company with deep pockets, hardware design expertise, and a reasonable certainty of return on investment.

Microsoft also had this problem when they tried to get manufacturers to be inventive with Windows 8/Windows RT; noone was prepared to take that financial risk, so Microsoft did it themselves with the Surface. I don't think it has resulted in much profit over the years.

sorry for the delayed response but I totally forgot about this discussion :P

I dont believe that what you say is accurate.

They surely had an idea of what a steambox(initial name I believe) should be, they wanted to make a PC for the living room (hence the controller that is supposed to retain the mouse like function/feeling for RTS games etc) and its main purpose was to be a console killer and lure console peasants to the art of PC gaming so that they can directly grab new customers from that virgin market of overpriced console gaming titles and treat them similarly as sony/ms/nintendo do using the PC exclusive genres and/or titles and the extra pc functions as a carrot for them to follow.

the problem (and that is my opinion) is that they didn't want to invest capital to RnD such a product and used their name and influence to drag companies that already are on the OEM business to do the work for them... but for reasons that I mentioned my previous posts those OEMs just missed the point.

SteamOS was the never the main purpose it was the means to make such a console killer viable (free OS/less licencing issues for apis/software/drivers etc)

If you look back at what Valve/Gaben said in 2012/2013, you will find that their ideas were quite broad, and not necessarily all that firm; they have produced a streaming device, but not produced the Valve-branded living-room version of Steambox they expected to make. For example, see https://www.theverge.com/2013/1/8/3852144/gabe-newell-interview-steam-box-future-of-gaming [External Link]

I would stand by my assessment of why neither they, nor other hardware OEMs, have managed to produce anything spectacular. It is, of course, just my opinion. Only the companies involved know what their dealings were.

From a personal viewpoint, the only thing that really matters to me is that they continue to invest in the Steam Client for Linux. Of their other ideas, the controller is far too small for my hands, the streaming link doesn't interest me, I prefer Ubuntu to SteamOS, and if I wanted a living-room box, I would build my own rather than buy - unless there is a sizeable hardware subsidy involved.

Perhaps they will eventually find a compelling recipe, but not yet.

View this comment - View article - View full comments

Quoting: elmapul

Quoting: etonbearsBest not to use UWP at all. Neither the architecture nor the market seem compelling to me.

we can avoid UWP all as we want, but how to convince developers to not use it?

Actually, I meant that comment from a developer perspective. The UWP development model is inferior to the Win32/64 model for desktop and laptop computers, where Microsoft are still reasonably strong. The whole point of UWP is to provide an equivalent to the iOS development model for tablets and phones, but that is a market in which Microsoft is almost invisible.

I'm not sure that Microsoft have yet realised that they are no longer important enough to enforce market adoption in all segments. It is quite possible that UWP will go the same way as Silverlight and countless other "innovative" technologies that were introduced more for Microsoft's benefit than for users or developers.

View this comment - View article - View full comments

Quoting: tuubi

Quoting: etonbearsIf a GPU manufacturer wanted to provide a custom graphics API ( like Vulkan ) they would write a UWP device app ( which has special permissions ) and the graphics API would be implemented against this device access app as a static library to be bound into any client UWP app ( UWP apps can't bind libraries dynamically other than the windows libraries ).

There's no whitelisting or enforced driver signing by Microsoft? I'd be very surprised. UWP is all about control after all. They want to be Apple so hard it hurts.

True, they do want control, but if a device manufacturer submits a UWP device app to allow direct hardware use they would risk regulatory scrutiny if they refuse to sign it. Usually they use commercial pressure with big companies.

Best not to use UWP at all. Neither the architecture nor the market seem compelling to me.

View this comment - View article - View full comments

Quoting: elmapul"Windows 10 is a general purpose computer OS, not locked to hardware. Both the CPU and GPU can be accessed through non-Microsoft APIs, even from within UWP programs; Vulkan APIs and drivers are supplied by the hardware manufacturers, not Microsoft."

i think that UWP is a sub set of apis, sandboxed and with some libraries.

it help to ensure either compatibility (you will not try to acess any hardware feature that is x86 exclusive), security (sandboxed)
subset (it wont support vulkan because it dont allow direct gpu acess to sendo vk commands)
but with some libraries to save developers time/money convincing then to use it and to make some effects without too much cpu/gpu usage.

example, you can use nvidia hair works to make hair, instead of coding it your self, it will save you time/money, at the cost of this single feature only working on nvidia hardware.
and since they know their hardare better than anyone else, they can optimize this feature better than you could if you try to do it by your self.
maybe they even have an asic on their gpu on part of it is only acessible if you use their api, so you cant compete in performance with ther oficial api.

i think that is what ms is trying to do here, but i'm not sure, i didnt read the documentation.

also i dont think xbox has even drivers to run vulkan or will ever have.

Yes, UWP is a subset - a very large subset - of the existing windows APIs. Most of the dropped APIs are to do with screen presentation in some way. The sandbox is limited in scope; while it does prevent a UWP application from invoking a device IOCTL directly, UWP has the "Device Access API" that a hardware manufacturer can use to provide a UWP application with access to their device.

If a GPU manufacturer wanted to provide a custom graphics API ( like Vulkan ) they would write a UWP device app ( which has special permissions ) and the graphics API would be implemented against this device access app as a static library to be bound into any client UWP app ( UWP apps can't bind libraries dynamically other than the windows libraries ).

What the performance would be like is anyone's guess, probably poor, since the entire idea of UWP is of questionable value to anyone but Microsoft; but it doesn't mean you can provide alternative graphics APIs if you really want to.

XBox, as I mentioned before, is a different case altogether.

View this comment - View article - View full comments

Quoting: elmapulUWP isnt compatible with Vulkan nor any program with opengl above 1.0
that means that vulkan apps cant run on xbox, hololens and whatever else microsoft put on the market.

currently lose the xbox market would do more harm to the pockets of the developers than losing mac/linux.

The XBOX is locked hardware with a restricted set of APIs. You would get better performance on XBOX by using the lowest level APIs rather than UWP.

Windows 10 is a general purpose computer OS, not locked to hardware. Both the CPU and GPU can be accessed through non-Microsoft APIs, even from within UWP programs; Vulkan APIs and drivers are supplied by the hardware manufacturers, not Microsoft.

I would not expect game makers to switch to UPW from XBOX low-level APIs, as it would reduce performance. If they want their game to support UWP on Windows 10, they will probably make the minimum changes necessary to existing DirectX code.

If they want to support additional platforms, they might do so by writing a Vulkan implementation, which could also be made to run on Windows 10 UWP. Whether that is a good idea or not depends on the game.

View this comment - View article - View full comments

@elmapul

"" " As soon as MS restrict Windows 10+ to running only signed applications"
why would they do that?
security? no one care about that otherwise they would use linux.
the ones who care use anti virus and think they are safe.

to make more profit?
they dont need to shoot then selfes in the foot, even if they are dumb enough to try that, their stake holders arent that dumb. ""

They would do that so they can take 30% of all application revenue on the windows platform, just like every other software distribution platform ( Steam , Apple store, XBox, PS etc ).

So, yes, to make more profit is why they would do it; their stakeholders are shareholders who only care about profit. If users have accepted Windows 10, they will accept store-locked applications. Anyone that doesn't will be free to look for an alternative.

"" "Windows 10 S(tupid) mode is testing the waters for a lock-out; "
nope it was an atempt to compete with the price of chromebooks. ""

Windows 10 S was an OS veriant for certain devices. Windows 10 S mode is the store-locked runtime mode that will be added to all all versions of Windows 10 ( Home, Pro and Enterprise ) in 2018/2019. According to Microsoft, users will still be allowed to switch to full windows mode to run non-store applications if they want.

History suggests Microsoft will incent their OEMs to ship devices with S mode enabled rather than full Windows mode. Seems like testing waters to me. :)

View this comment - View article - View full comments

Quoting: tonR

Quoting: NeverthelessI don't think it ever was a device species at all. I don't know what Valve thought of it when they released it, but for me it was a Linux PC with a controller for the living room. That concept is not dead, nor are the Steam Machines sold. Nothing has failed here, because nothing gets abandoned. Owners of a Steam Machine still have got a Linux PC for the living room and 4000+ games to play.

That's my question. What was/is in Valve's people mind when they launched Steam Machine? To promoted Linux? For fun?

I never said failed. Yes, Valve is still making money, Steam Machine or not, but every businesses in this world wanted to maximizing profit to even every single cent. So I'm very curious about the reason of Steam Machine existance.

I don't think there was a clear idea, on Valve's part, what a Steam Machine was for. It was a strand in the "open" gaming platform they wanted to exist, just as SteamOS and the Linux Steam runtime are strands of that platform.

My recollection is that they stated that they wished to foster hardware innovation, so there was no fixed/clear definition of a Steam Machine, but that they expected novel and upgradeable hardware offerings ( not just living room PCs ).

Unfortunately, that is difficult to achieve. PCs are already, de facto, flexible and upgradeable, so what can be added to that? Novel hardware that might have superior performance or price characteristics would take the form of custom designed components; but that requires a company with deep pockets, hardware design expertise, and a reasonable certainty of return on investment.

Microsoft also had this problem when they tried to get manufacturers to be inventive with Windows 8/Windows RT; noone was prepared to take that financial risk, so Microsoft did it themselves with the Surface. I don't think it has resulted in much profit over the years.

View this comment - View article - View full comments

Quoting: tonRI'm not surprise that Steam Machine may discontinued. Here is very simple question, who is the Valve's target customers for Steam Machine?
- Streamer (legally or not)? Kodi box is cheaper.

- PC gamers? Most of them build their own rig; more powerful and more value for money.

- Console gamers? Exclusive (or timed exclusive) is a key here. I'm remember GoL readers totally disagree with me when I suggest CDPR should sell their game on GOG exclusively.

- "PC as truck" (like late Steve Jobs comment about post-pc era.)? There are more old-school advanced "truck" called laptop.

- People who want try Linux?

- Steam Machine as "benchmark standard device" (like iPhone, one spec hardware) for developers? It requires locking- down/out the device. The problem is Linux is GPL-open. (And that's why Google currently experimenting Fuchsia)

Yes, I am sure you are right about cutting new supply. It will certainly work with many new computer users.

Just to reiterate, UWP does not affect the majority of WIN32 and COM. It's more like the move from X servers to Wayland plus signed binaries, as I understand it.

View this comment - View article - View full comments

Quoting: Leopard

Quoting: etonbears

Quoting: NeverthelessLinux wasn't ready at the time Valve announced Steam Machines and SteamOS. It stll isn't quite ready. AMD drivers need time, VR needs time, Vulkan still needs time.
I don't know why Valve came out with it prematurely, but I guess something had to be done at that time, to push Linux development and to slow down Microsoft closing down Windows. Microsoft is doing so now. They are clearly pushing UWP and Windows S-Mode and show the tendency to want to abandon Win32, which Steam needs to exist on Windows. According to his latest statements, at least G. Newell seems to see it that way. So in my opinion abandoning Linux would be an idiotic thing to do by Valve, and the only question is how they plan to push it, and when they think the time is right!

The main driver for Valve has always been the potential for a future Windows to only allow Microsoft signed applications to run, and for Midcosoft to only sign applications that are sold through, and conform to the policies of, the windows store ( copying the Apple iOS model ).

Almost all of the WIN32 API remains in Windows 10, and will continue to exist. WIN32 and COM are the key, lowest level, building blocks for all of Microsoft's other APIs, including UWP. The real issue isn't technical ( Valve could easily produce a UWP Steam Client ), but a matter of commercial control, as they would not get a signed UWP version of the Steam Client without agreeing to give Microsoft a percentage of any revenue gained through it.

As soon as MS restrict Windows 10+ to running only signed applications from the Windows store, Steam for Windows is toast, probably along with GOG, Humble and anyone else that acts as a store-front for Windows applications.

And that is the critical point. Microsoft won't do that until they have a better store and near title equality in it.

They have most games already, since their strategy seems to be to blur the distinction between developing for PC or XBox through common APIs ( develop for XBox, and recompile to run under UWP ). They are very good at providing the tools developers need to support their platforms, and always have been.

Windows 10 S(tupid) mode is testing the waters for a lock-out; it will come when MS feel it will be largely accepted, I guess. I'm sure there will be legal challenges, but that will be too slow for companies put out of business.

View this comment - View article - View full comments